Interest Rate Prediction:

Data Cleaning:

Data Exploration:

From the above results, we can observe the correlation of each pf the predictors with the target variable;
'paid_interest' is highly positively correlated variable with the target variable (interest_rate), whereas
'total_debit_limit' is highly negatively correlated variable with the target variable.

From the above scatte, the correlation with the target variable is evident. For lower interest rates, the total_debit_limit is high as compared to those for higher interest rates. Similarly, for lower interest rates the paid interest is also low and as the rate increase the paid_interest value also increases.

Ofcourse, there are exceptions to this analysis, but that can be justified by the moderate correlation value.

Final Clean Data Set:

Data Transformation:

Data Split:

Decision Tree Regressor:

Clearly, the model is overfitting on the train set, hence we tune the hyperparameters and as per the results develp a new model

Considering the 0.99 r2_score of the parameter tuned DTR model, the above interactive scatter plot explains the linear regression between the actual test set and the predicted interest rate values.

Random Forest Regressor:

Considering the 0.99 r2_score of the parameter tuned RFR model, the above interactive scatter plot explains the linear regression between the actual test set and the predicted interest rate values.